Parsing HTML in python – lxml or BeautifulSoup? Which of these is better for what kinds of purposes?
Question:
From what I can make out, the two main HTML parsing libraries in Python are lxml and BeautifulSoup. I’ve chosen BeautifulSoup for a project I’m working on, but I chose it for no particular reason other than finding the syntax a bit easier to learn and understand. But I see a lot of people seem to favour lxml and I’ve heard that lxml is faster.
So I’m wondering what are the advantages of one over the other? When would I want to use lxml and when would I be better off using BeautifulSoup? Are there any other libraries worth considering?
Answers:
For starters, BeautifulSoup is no longer actively maintained, and the author even recommends alternatives such as lxml.
Quoting from the linked page:
Version 3.1.0 of Beautiful Soup does
significantly worse on real-world HTML
than version 3.0.8 does. The most
common problems are handling
tags incorrectly, “malformed start
tag” errors, and “bad end tag” errors.
This page explains what happened, how
the problem will be addressed, and
what you can do right now.
This page was originally written in
March 2009. Since then, the 3.2 series
has been released, replacing the 3.1
series, and development of the 4.x
series has gotten underway. This page
will remain up for historical
purposes.
tl;dr
Use 3.2.0 instead.
I’ve used lxml with great success for parsing HTML. It seems to do a good job of handling “soupy” HTML, too. I’d highly recommend it.
Here’s a quick test I had lying around to try handling of some ugly HTML:
import unittest
from StringIO import StringIO
from lxml import etree
class TestLxmlStuff(unittest.TestCase):
bad_html = """
<html>
<head><title>Test!</title></head>
<body>
<h1>Here's a heading
<p>Here's some text
<p>And some more text
<b>Bold!</b></i>
<table>
<tr>row
<tr><td>test1
<td>test2
</tr>
<tr>
<td colspan=2>spanning two
</table>
</body>
</html>"""
def test_soup(self):
"""Test lxml's parsing of really bad HTML"""
parser = etree.HTMLParser()
tree = etree.parse(StringIO(self.bad_html), parser)
self.assertEqual(len(tree.xpath('//tr')), 3)
self.assertEqual(len(tree.xpath('//td')), 3)
self.assertEqual(len(tree.xpath('//i')), 0)
#print(etree.tostring(tree.getroot(), pretty_print=False, method="html"))
if __name__ == '__main__':
unittest.main()
Don’t use BeautifulSoup, use
lxml.soupparser then you’re sitting on top of the power of lxml and can use the good bits of BeautifulSoup which is to deal with really broken and crappy HTML.
Pyquery provides the jQuery selector interface to Python (using lxml under the hood).
http://pypi.python.org/pypi/pyquery
It’s really awesome, I don’t use anything else anymore.
A somewhat outdated speed comparison can be found here, which clearly recommends lxml, as the speed differences seem drastic.
In summary, lxml is positioned as a lightning-fast production-quality html and xml parser that, by the way, also includes a soupparser module to fall back on BeautifulSoup’s functionality. BeautifulSoup is a one-person project, designed to save you time to quickly extract data out of poorly-formed html or xml.
lxml documentation says that both parsers have advantages and disadvantages. For this reason, lxml provides a soupparser so you can switch back and forth. Quoting,
BeautifulSoup uses a different parsing approach. It is not a real HTML
parser but uses regular expressions to dive through tag soup. It is
therefore more forgiving in some cases and less good in others. It is
not uncommon that lxml/libxml2 parses and fixes broken HTML better,
but BeautifulSoup has superiour support for encoding detection. It
very much depends on the input which parser works better.
In the end they are saying,
The downside of using this parser is that it is much slower than
the HTML parser of lxml. So if performance matters, you might want
to consider using soupparser only as a fallback for certain cases.
If I understand them correctly, it means that the soup parser is more robust — it can deal with a “soup” of malformed tags by using regular expressions — whereas lxml is more straightforward and just parses things and builds a tree as you would expect. I assume it also applies to BeautifulSoup itself, not just to the soupparser for lxml.
They also show how to benefit from BeautifulSoup‘s encoding detection, while still parsing quickly with lxml:
>>> from BeautifulSoup import UnicodeDammit
>>> def decode_html(html_string):
... converted = UnicodeDammit(html_string, isHTML=True)
... if not converted.unicode:
... raise UnicodeDecodeError(
... "Failed to detect encoding, tried [%s]",
... ', '.join(converted.triedEncodings))
... # print converted.originalEncoding
... return converted.unicode
>>> root = lxml.html.fromstring(decode_html(tag_soup))
(Same source: http://lxml.de/elementsoup.html).
In words of BeautifulSoup‘s creator,
That’s it! Have fun! I wrote Beautiful Soup to save everybody time.
Once you get used to it, you should be able to wrangle data out of
poorly-designed websites in just a few minutes. Send me email if you
have any comments, run into problems, or want me to know about your
project that uses Beautiful Soup.
--Leonard
Quoted from the Beautiful Soup documentation.
I hope this is now clear. The soup is a brilliant one-person project designed to save you time to extract data out of poorly-designed websites. The goal is to save you time right now, to get the job done, not necessarily to save you time in the long term, and definitely not to optimize the performance of your software.
Also, from the lxml website,
lxml has been downloaded from the Python Package Index more than two
million times and is also available directly in many package
distributions, e.g. for Linux or MacOS-X.
And, from Why lxml?,
The C libraries libxml2 and libxslt have huge benefits:…
Standards-compliant… Full-featured… fast. fast! FAST! … lxml
is a new Python binding for libxml2 and libxslt…
For sure i would use EHP. It is faster than lxml, much more elegant and simpler to use.
Check out. https://github.com/iogf/ehp
<body ><em > foo <font color="red" ></font></em></body>
from ehp import *
data = '''<html> <body> <em> Hello world. </em> </body> </html>'''
html = Html()
dom = html.feed(data)
for ind in dom.find('em'):
print ind.text()
Output:
Hello world.
From what I can make out, the two main HTML parsing libraries in Python are lxml and BeautifulSoup. I’ve chosen BeautifulSoup for a project I’m working on, but I chose it for no particular reason other than finding the syntax a bit easier to learn and understand. But I see a lot of people seem to favour lxml and I’ve heard that lxml is faster.
So I’m wondering what are the advantages of one over the other? When would I want to use lxml and when would I be better off using BeautifulSoup? Are there any other libraries worth considering?
For starters, BeautifulSoup is no longer actively maintained, and the author even recommends alternatives such as lxml.
Quoting from the linked page:
Version 3.1.0 of Beautiful Soup does
significantly worse on real-world HTML
than version 3.0.8 does. The most
common problems are handling
tags incorrectly, “malformed start
tag” errors, and “bad end tag” errors.
This page explains what happened, how
the problem will be addressed, and
what you can do right now.This page was originally written in
March 2009. Since then, the 3.2 series
has been released, replacing the 3.1
series, and development of the 4.x
series has gotten underway. This page
will remain up for historical
purposes.tl;dr
Use 3.2.0 instead.
I’ve used lxml with great success for parsing HTML. It seems to do a good job of handling “soupy” HTML, too. I’d highly recommend it.
Here’s a quick test I had lying around to try handling of some ugly HTML:
import unittest
from StringIO import StringIO
from lxml import etree
class TestLxmlStuff(unittest.TestCase):
bad_html = """
<html>
<head><title>Test!</title></head>
<body>
<h1>Here's a heading
<p>Here's some text
<p>And some more text
<b>Bold!</b></i>
<table>
<tr>row
<tr><td>test1
<td>test2
</tr>
<tr>
<td colspan=2>spanning two
</table>
</body>
</html>"""
def test_soup(self):
"""Test lxml's parsing of really bad HTML"""
parser = etree.HTMLParser()
tree = etree.parse(StringIO(self.bad_html), parser)
self.assertEqual(len(tree.xpath('//tr')), 3)
self.assertEqual(len(tree.xpath('//td')), 3)
self.assertEqual(len(tree.xpath('//i')), 0)
#print(etree.tostring(tree.getroot(), pretty_print=False, method="html"))
if __name__ == '__main__':
unittest.main()
Don’t use BeautifulSoup, use
lxml.soupparser then you’re sitting on top of the power of lxml and can use the good bits of BeautifulSoup which is to deal with really broken and crappy HTML.
Pyquery provides the jQuery selector interface to Python (using lxml under the hood).
http://pypi.python.org/pypi/pyquery
It’s really awesome, I don’t use anything else anymore.
A somewhat outdated speed comparison can be found here, which clearly recommends lxml, as the speed differences seem drastic.
In summary, lxml is positioned as a lightning-fast production-quality html and xml parser that, by the way, also includes a soupparser module to fall back on BeautifulSoup’s functionality. BeautifulSoup is a one-person project, designed to save you time to quickly extract data out of poorly-formed html or xml.
lxml documentation says that both parsers have advantages and disadvantages. For this reason, lxml provides a soupparser so you can switch back and forth. Quoting,
BeautifulSoup uses a different parsing approach. It is not a real HTML
parser but uses regular expressions to dive through tag soup. It is
therefore more forgiving in some cases and less good in others. It is
not uncommon that lxml/libxml2 parses and fixes broken HTML better,
but BeautifulSoup has superiour support for encoding detection. It
very much depends on the input which parser works better.
In the end they are saying,
The downside of using this parser is that it is much slower than
the HTML parser of lxml. So if performance matters, you might want
to consider using soupparser only as a fallback for certain cases.
If I understand them correctly, it means that the soup parser is more robust — it can deal with a “soup” of malformed tags by using regular expressions — whereas lxml is more straightforward and just parses things and builds a tree as you would expect. I assume it also applies to BeautifulSoup itself, not just to the soupparser for lxml.
They also show how to benefit from BeautifulSoup‘s encoding detection, while still parsing quickly with lxml:
>>> from BeautifulSoup import UnicodeDammit
>>> def decode_html(html_string):
... converted = UnicodeDammit(html_string, isHTML=True)
... if not converted.unicode:
... raise UnicodeDecodeError(
... "Failed to detect encoding, tried [%s]",
... ', '.join(converted.triedEncodings))
... # print converted.originalEncoding
... return converted.unicode
>>> root = lxml.html.fromstring(decode_html(tag_soup))
(Same source: http://lxml.de/elementsoup.html).
In words of BeautifulSoup‘s creator,
That’s it! Have fun! I wrote Beautiful Soup to save everybody time.
Once you get used to it, you should be able to wrangle data out of
poorly-designed websites in just a few minutes. Send me email if you
have any comments, run into problems, or want me to know about your
project that uses Beautiful Soup.--Leonard
Quoted from the Beautiful Soup documentation.
I hope this is now clear. The soup is a brilliant one-person project designed to save you time to extract data out of poorly-designed websites. The goal is to save you time right now, to get the job done, not necessarily to save you time in the long term, and definitely not to optimize the performance of your software.
Also, from the lxml website,
lxml has been downloaded from the Python Package Index more than two
million times and is also available directly in many package
distributions, e.g. for Linux or MacOS-X.
And, from Why lxml?,
The C libraries libxml2 and libxslt have huge benefits:…
Standards-compliant… Full-featured… fast. fast! FAST! … lxml
is a new Python binding for libxml2 and libxslt…
For sure i would use EHP. It is faster than lxml, much more elegant and simpler to use.
Check out. https://github.com/iogf/ehp
<body ><em > foo <font color="red" ></font></em></body>
from ehp import *
data = '''<html> <body> <em> Hello world. </em> </body> </html>'''
html = Html()
dom = html.feed(data)
for ind in dom.find('em'):
print ind.text()
Output:
Hello world.